Product Analysis with Cursor and Claude Code

I’ve started using Cursor + Claude Code for my product analysis, and it’s slowly changing how I approach product iteration. So far, I like the experience! Here's what I've learned 🧵

First of all, we all know a true “source of truth” is a myth in most companies. Data is spread across different platforms, and it requires some stitching to gain deeper insight. Before, my workflow looked like this:
- Event analytics in one tab
- DB queries in another
- Organizing datasets and creating visualizations with ChatGPT
- Writing a report in Notion or Google Docs
- Discussing with the team and deciding next steps for iteration

It worked, but it felt like I was reinventing the wheel every time.
Now I make everything happen in one place:
- I open Cursor
- Claude Code pulls data from BigQuery MCP + event APIs
- Claude Code analyzes it and drafts a markdown report directly in Cursor
- I iterate on the doc by asking Claude more questions
- When action items become clear, Claude creates Linear tickets or fixes issues directly in the codebase
- Every analysis becomes reusable historical context on my machine

This workflow moved the entire “Build → Measure → Learn → Build” cycle under one roof. I’ve just begun to explore this workflow, but I’m curious about how it can evolve in my day-to-day job.